Research entry
ACTIVE
Verification receipts for coding agents
Engineering research into making agent work auditable through independent verification receipts.
Research question
Does an independent, execution-backed verification plane reduce unjustified confidence in agent-authored changes compared to narrative self-reports?
Hypothesis
Durable receipts from builds, tests, lints, and API-drift checks create a review surface that resists fluent but unverified patches.
Methodology
Build Aftermath as an independent examiner. Pair with Stackglass evidence collection. Evaluate qualitatively on real repositories during agent workflows.
Results
Tooling exists and is usable. Controlled user studies and quantified defect-catch rates are not claimed on this site.
Limitations
No formal study population; results are engineering observations.